atus, in his account of this experiment, as he generally is in other cases, and therefore probably wrote his account of it from his memory only.

Much has been said on this experiment; and it is thought very extraordinary, that a man of Sir Isaac's accurate attention should have overlooked a circumstance, the effect of which now appears to be so considerable. But it has happily occurred to Mr Michel, that, as Sir Isaac Newton observes, he used to put saccharum saturni into his water to increase its refractive power; the lead, even in this form, might increase the dissipative refraction, as it does in the composition of glass; and if so, that this would account for Newton's not finding his dissipative power of water less than that of the glass prisms, which he otherwise ought to have done, if he had tried the experiment as he said he did.

Accordingly he included a prism of glass in water, as highly impregnated with saccharum saturni as it would bear, the proportion of saccharum to water being about as 5 to 11. When the image, seen through the water (so impregnated) and a glass prism, was in its natural place, it still was coloured, though very little: he thought not more than a fourth part as much as when seen through the plain water, and the prism in its natural place; so that he had no doubt, but that, if his prism had had a little less of the dispersing power, its errors would have been perfectly corrected."

Besides the experiments of Mr Delaval above related, and which were made on the colours of transparent bodies, he has lately published an account of some made upon the permanent colours of opaque substances; the discovery of which must be of the utmost consequence in the arts of colour-making and dyeing. These arts, he observes, were in very remote ages carried to the utmost height of perfection in the countries of Phoenicia, Egypt, Palestine, India, &c., and that the inhabitants of these countries also excelled in the art of imitating gems, and tinging glass and enamel of various colours. The colours used in very ancient paintings were as various as those now in use, and greatly superior both in beauty and durability. The paints used by Apelles were so bright, that he was obliged to glaze his pictures with a dark coloured varnish, lest the eye should be offended by their excessive brightness; and even these were inferior to what had been used among the ancient Egyptians. Pliny complains that the art of painting was greatly decayed in his time; and the moderns were not furnished with any means of retrieving the art, until they began to avail themselves of experimental observations.

The changes of colour in permanently coloured bodies, our author observes, are produced by the same laws which take place in transparent colourless substances; and the experiments by which they can be investigated consist chiefly of various methods of uniting the colouring particles into larger, or dividing them into smaller masses. Sir Isaac Newton made his experiments chiefly on transparent substances; and in the few places where he treats of others, acknowledges his deficiency of experiments. He makes the following remark, however, on those bodies which reflect one kind of light and transmit another, viz. that "If these glasses or liquors were so thick and massy that no light could get through them, he questioned whether they would not, like other opaque bodies, appear of one and the same colour in all positions of the eye; though he could not yet affirm it from experience." It was the opinion of this great philosopher, that all coloured matter reflects the rays of light, some reflecting the more refrangible, and others the less refrangible rays more copiously; and that this is not only a true reason of these colours, but likewise the only reason. He was likewise of opinion, that opaque bodies reflect the light from their interior surface by some power of the body evenly diffused over and external to it. With regard to transparent-coloured liquors, he expresses himself in the following manner: "A transparent body, which looks of any colour by transmitted light, may also look of the same colour by reflected light; the light of that colour being reflected by the farther surface of that body, or by the air beyond it: and then the reflected colour will be diminished, and perhaps cease, by making the body very thick, and pitching it on the back side to diminish the reflection of its farther surface, so that the light reflected from the tinging particles may predominate. In such cases, the colour of the reflected light will be apt to vary from that of the light transmitted.

To investigate the truth of these opinions, Mr Delaval entered upon a course of experiments with transparent coloured liquors and glasses, as well as with opaque and semi-transparent bodies. From these he discovered several remarkable properties of the colouring matter; particularly, that in transparent coloured substances it does not reflect any light; and when, by intercepting the light which was transmitted, it is hindered from passing through such substances, they do not vary from their former colour to any other, but become entirely black (A).

This incapacity of the colouring particles of transparent

(A) Here our author observes, that he makes use of the word colour only to express those called primary; such parent bodies to reflect light, being deduced from very numerous experiments, may therefore be held as a general law. It will appear the more extensive, if we consider, that, for the most part, the tinging particles of liquors or other transparent substances are extracted from opaque bodies; that the opaque bodies owe their colours to those particles, in like manner as the transparent substances do; and that by the loss of them they are deprived of their colours.

For making his experiments, Mr Delaval used small phials of flint-glass, whose form was a parallelopiped, and their height, exclusive of the neck, about two inches, the base about an inch square, and the neck two inches in length. The bottom and three sides of each of these phials were covered with a black varnish; the cylindrical neck, and the anterior side, except at its edges, being left uncovered. He was careful to avoid any crevices in the varnish, that no light might be admitted except through the neck or anterior side of the phials.

In these experiments it is of importance to have the phials perfectly clean; and as many of the liquors are apt to deposit a sediment, they ought to be put into the phials only at the time the experiments are to be made. The uncovered side of the phials should not be placed opposite to the window through which the light is admitted; because in that situation the light would be reflected from the farther side of the phial; and our author observes, that smooth black substances reflect light very powerfully. But as it is a principal object in the experiment, that no light be transmitted through the liquor, this is best accomplished by placing the uncovered side of the phial in such a situation that it may form a right angle with the window.

With these precautions, our author viewed a great number of solutions, both of coloured metallic salts and of the tinging matter of vegetables; universally observing, that the colour by reflection was black, whatever it might be when viewed by transmitted light. If these liquors, however, are spread thin upon any white ground, they appear of the same colour as when viewed by transmitted light; but on a black ground they afford no colour, unless the black body be polished; in which case the reflection of the light through it produces the same effect as transmission.

The experiments with tinged glasses were in many respects analogous to those with transparent-coloured liquors. For these he made several parcels of colourless glasses, principally using one composed of equal parts of borax and white sand. The glass was reduced to powder, and afterwards ground, together with the ingredients by which the colours were imparted. "This method (says he) of incorporating the tinging particles is greatly preferable to mixing them with the raw materials; and the glasses thus composed excel most others in hardness, being scarcely inferior in lustre to real gems."

The result of all the experiments made in this manner was, that when matter is of such thinness, and the tinge so diluted, that light can be transmitted through it, the glasses then appear vividly coloured; but when they are in larger masses, and the tinging matter is more densely diffused through them, they appear black; for these, as well as the transparent-coloured liquors, show their colour by transmission. The following experiments were made with a view to determine the proportion of tinging matter which produces colour or blackness.

1. Glass was tinged green by adding to it $\frac{1}{30}$th of its weight of copper; and that whether the latter was agents to used in its metallic or calcined state.

2. A blue glass was made by the addition of zaffire, the portion of a purple one by manganese, a red glass by gold, and tinging yellow glasses by silver and calcined iron. A yellow matter glass resembling a topaz was likewise made by the addition of a small quantity of charcoal in powder. The same colour was likewise procured by the addition of wheat-flour, rosin, and several other inflammable matters. Small pieces of each of these glasses being ground by a lapidary, resembled gems of their different colours.

3. Having formed pieces of such glasses about two inches thick, he inclosed them in black cloth on all sides, except their farther and anterior surfaces. In this situation each of them showed a vivid colour when light was transmitted through them; but when the posterior surface was likewise covered with the cloth to prevent this transmission, no other colour than black was exhibited by any of them.

4. When plates of transparent-coloured glass, somewhat thicker than common window-glass, were made use of, they always exhibited their colours by transmitted light.

5. On intercepting the light transmitted through these coloured plates, they as constantly appeared black when placed in such a direction as to form a right angle with the window.

From these phenomena Mr Delaval deduced the following observations: 1. That the colouring particles do not reflect any light. 2. That a medium, such as Sir Isaac Newton has described, is diffused over both the anterior and farther surfaces of the plates, whereby objects are equally and regularly reflected as by a mirror. Hence, when it is said that light is reflected by the surface of any substance, it should be understood from this expression, that the reflection is effected by the medium diffused over its surface.

6. When a lighted candle is placed near one of those on the coloured plates, the flame is reflected by the medium section of which is diffused over the anterior surface. The image the light of thus reflected entirely resembles the flame in size and coloured colour; being scarcely diminished, and not in the least tinged by the coloured glass.

7. If the plate be not so intensely coloured, or so massy, as to hinder the transmission of the light of the candle, there appears a secondary image of the flame, which is reflected by the medium contiguous to the farther surface of the glass; and as the light thus reflected passes through the coloured glass, it is tinged very vividly.

8. When such a mixture of them as does not compose whiteness, or any of the gradations between white and black; such as are called by Sir Isaac Newton, gray, dun, or russet brown. 8. When the glass used in this experiment is of a green colour, the image of the flame is always of a bright green; and when glasses of other colours are used, that of the secondary flame is always the same with that of the glass.

9. The secondary image is less than that reflected from the anterior surface. This diminution is occasioned by the loss of that part of the light which is absorbed in passing through the coloured glass. For whenever any medium transmits one sort of rays more copiously than the rest, it stops a great part of the differently coloured rays. Much more light also is lost in passing through coloured than transparent substances. In making these observations, it is proper to choose coloured plates of glass which are not in every part of an equal thickness, that the secondary image may not coincide with that reflected from the anterior surface, and be intercepted by it.

10. When the plates are so thick, and so copiously coloured, that the light cannot penetrate to their farther surface, they appear intensely black in whatever direction they are viewed, and afford no secondary image, but only reflect, from their anterior surface, the flame, or any other objects that are opposed to them. These objects are represented in their own proper colours, and are as free from tinge as those reflected from quicksilvered glass, or specula made of white metals.

Hence again it is manifest, that the colouring particles do not possess any share of reflective power; for if they had any share in this reflection, they would certainly impart some share of colour to the light they reflected. Hence also it appears, that transparent coloured bodies, in a solid state, possess no more reflective power than those in a fluid state.

Our author next considers the colouring particles themselves, pure, and unmixed with other media. In order to procure masses made up of such particles, several transparent coloured liquors were reduced to a solid consistence by evaporation. By employing a gentle heat, the colouring matter may thus remain unimpaired; and is capable of having its particles again separated by water or other liquids, and tinging them as before.

In this state the colouring particles reflect no light, and therefore appear uniformly black, whatever substance they have been extracted from. In the course of his experiments, Mr Delaval made use of the infusions of brazil wood, logwood, fustic, turmeric, red saunders, alkanet, sap-green, kermes, and all the other transparent coloured liquors he had tried before, among which were infusions of red and yellow flowers, without observing the least variation in the result.

Some liquors are apt to become totally opaque by evaporation; the reason of which may be the crystallization of saline matters, or the coalescence of the particles into masses, differing considerably in density from the menstrus in which they were dissolved. When this opacity takes place, our author has constantly observed, that they became incapable of entering the pores of wool, silk, or other matters of that kind, or of adhering to their surface; and consequently unfit for the purposes of dyeing. This he supposes to arise from their increased bulk; for the attractive force by which the particles cohere together is weakened in proportion as their bulk increases; so that the degree of magnitude of the colouring particles, which is essential to the capacity of liquors, is inconsistent with the minuteness requisite for dyeing. An instance of this is given in an infusion of fustic. Having infused some of this wood in such a quantity of water, that the latter was saturated with the colouring particles, he evaporated the liquor to a solid consistence, with an uninterrupted but very gentle heat. During every part of the process the liquor continued transparent, and the solid extract yielded by it transmitted a yellow colour when spread thin, but appeared black when thicker masses were viewed. Having prepared another pint of this liquor, he evaporated half the water, and allowed the remainder to become cold. In this state it became turbid and opaque; on filtering, a transparent tincture passed through, an opaque fecula remaining on the paper. This fecula did not adhere to the paper, but was easily separable from it: on being dried, it appeared white with a slight tinge of yellow; but was nevertheless soluble in water, and by solution gave a liquid in all respects similar to the original infusion.

"From these circumstances (says he) it appears that a given proportion of water, or a sufficient degree of heat, is requisite to the solution of the colouring particles of fustic. And experience evinces that those particles which are too gross to pass through filtering paper, are incapable of entering the pores, or firmly cohering to the surface of bodies. Many ingredients, such as the colouring particles of logwood, kermes, and various other matters, are soluble in water in every proportion; and therefore their infusions are not subject to become opaque or turbid during their evaporation. The solid extracts obtained by evaporation reflect no colour, but are black."

Our author also formed solid masses by mixing a small quantity of drying oil with pigments which consist chiefly of colouring matter; as Prussian blue, indigo, and sap green. These paints likewise exhibit their respective colours only by transmitted light, appearing entirely black when viewed by reflection. Instances of blackness arising from this density of the colouring matter, may be observed in several kinds of fruits, as black currants, cherries, &c. for the juices of these appear red when spread thin on a white ground, or otherwise viewed by transmitted light.

Mr Delaval's next attempt was to consider the action and properties of the colouring particles of opaque bodies themselves, and the means by which these colours are produced. Here our author endeavours to prove, that these colours of opaque bodies appear on the same principles as those already mentioned, which seem black when very dense, but show their proper tinge when spread thin upon a white ground. On this subject the following experiments were made:

1. Grass, and other green leaves of plants, were digested in rectified spirit of wine; by which means a transparent green tincture was obtained. One of the phials formerly mentioned being filled with this liquid, it was observed to transmit a vivid green colour; but the other part of the tincture, which was contiguous to the uncovered side of the phial, reflected no light, and therefore appeared black.

2. Having poured some of the tincture into a China cup, the bottom was thereby made to look green, exactly. actly resembling the colour which had been extracted from the leaves.

3. After the colour had been totally extracted by the visous spirit, the leaves remained apparently unaltered, either as to figure or texture; but were entirely white, or had their whiteness slightly tinged with brown.

4. Red, purple, and blue flowers, were also digested in spirit of wine, all of which yielded their colouring matter to the spirit, and became white by being deprived of it. From most of these flowers, however, the spirit acquired either no tinge at all, or only a very faint one; but when acidulated, it became red, and by the addition of an alkali appeared blue, purple, or green, according to the quantity of alkali, and the nature of the infusion. In these states, all of them, when viewed by transmitted light, or poured upon a white ground, showed their colours, but universally appeared black by reflexion.

5. Red, purple, and blue flowers, were digested in water slightly acidulated with nitrous acid. Thus, red infusions were obtained, which, by saturation with sea-salt, might be preserved for many years.

6. The same liquors were changed green, blue, or purple, by the addition of an alkali: but here the case was the same as before; all of them yielding vivid colours by transmission, but none by reflection. In making this experiment, care must be taken to add the alkali very gradually; for if too much is put in at once to the red liquor, the immediate colours between the red and the green will be wanting. To half an ounce of the red infusion it is proper to add, at once, only the smallest quantity that can be taken upon the point of a pen; repeating this addition slowly, until each of the colours be produced.

7. The flowers, after having been repeatedly macerated in acidulated water, lost their colouring matter, and became white.

8. Yellow flowers also communicated their colours to water and to spirit of wine. The infusion and tinctures of these flowers were subjected to the same experiments as had been employed in the examination of the liquors already mentioned; and appeared yellow by transmitted light, but did not reflect any colour.

9. White paper, linen, &c., may be tinged of any of these colours, by dipping them in the infusions; and the consideration of the manner in which the colours are imparted to the linen, affords much insight into the manner in which natural colours are produced. It has already been observed, that, when the colouring matter of plants is extracted from them, the solid fibrous parts, thus divested of their covering, display their natural whiteness. White linen, paper, &c., are formed of such fibrous vegetable matter; which is bleached by dissolving and detaching the heterogeneous colouring particles. When these are dyed or painted with vegetable colours, it is evident that they do not differ in their manner of acting on the rays of light from natural vegetable bodies; both yielding their colours by transmitting, through the transparent coloured matter, the light which is reflected from the white ground. This white matter frequently exists, without any considerable mixture, in plants, while they are in a state of vegetation; as cotton, white flowers, the pith, wood, seeds, roots, and other parts of several kinds of vegetables. When decayed trees, &c., have been long exposed to the atmosphere, their coloured juices are sometimes so perfectly extracted, that the fibres appear white. This white matter is not distinct from the vegetable earth to which plants are reduced by burning. Mr. Delaval has rendered ashes intensely white, by carefully calcining them, and afterwards grinding with a small proportion of nitre, and exposing them to such a degree of intensely heat as would cause the nitre delagrate with the remaining quantity of phlogiston. Lastly, the ashes were digested with muriatic acid, in order to dissolve the ferruginous matter diffused through them, and repeatedly washing the remainder in water. Mixing ashes thus purified with borax, and applying a vitrifying heat, an opaque enamel is obtained, remarkable for its whiteness.

Hence it appears, that the earth which forms the substance of plants is white, and separable from that earth of substance which gives to earth its peculiar colour; that plants, whenever it is pure and unmixed, or diffused through colourless media, it shows its native whiteness; and that the only vegetable matter endowed with a reflective reflect power. It may be discovered, however, by other light means than that of burning; thus, roses may be whitened by exposing them to the vapour of burning sulphur: an effect which cannot be attributed to the sulphuric acid, but to the phlogiston contained in that vapour. This was proved to be the case, by exposing several kinds of red and purple flowers to the phlogistic vapour issuing from hepar sulphuris; and by this every one of them was whitened; their colour being afterwards restored by the addition of an acid either mineral or vegetable.

Thus (says Mr. Delaval) it appears, that the colouring matter of the flowers is not discharged or re-matter removed, but only dissolved by carbonic acid; and thereby divided into particles too minute to exhibit any colour. In this state, together with the vegetable juice in which they are diffused, they form a colourless transparent covering, through which the white matter of the flowers is seen untinged. The colouring particles of plants consist principally of inflammable matter, and their solubility in carbonic acid, and union with it, are analogous to the action of other inflammable bodies upon each other. Thus, ether dissolves all essential and expressed oils, animal empyreumatic oils and resins. Sulphur, camphor, and almost all substances abounding in phlogiston, are soluble in oils, ardent spirits, or other inflammable menstrua. The manner in which the red colour of vegetable flowers is restored, appears to be explicable from known chemical laws. When acids are applied to the whitened flowers, they unite with the phlogiston which the sulphur had communicated, and disengage it from the colouring particles; which, being thus extricated, resume their original magnitude and hue. A change of the same kind is also produced by fixed alkali, which, like the acids, has a strong attraction for phlogiston, always changes the whitened flowers to a blue, purple, or green colour.

In like manner, the action of the rays of light operates upon coloured bodies. Thus, dyed silk, or stroyed other substances of that kind, when exposed to the light of the sun's sun. sun's light, are deprived of their colour in every part on which the rays are allowed to act; whilst those preserve their colour which are defended from the light by the folds of the cloth, or intervention of any opaque body. The colours, thus impaired, may be restored, if acids are applied while the injury is recent; but they are afterwards apt to fly off, on account of that volatility which is constantly imparted by inflammable matter to any other with which it is united."

Our author now proceeds, at considerable length, to prove the identity of the solar light and carbuncle acid; but as recent experiments have shown that these two are essentially distinct, we omit his argumentation upon this head. The error of his theory in this respect, however, does not in the least affect the doctrine concerning colours above laid down: on the contrary, the latest experiments have determined, that carbuncle acid in its grossest form, viz. that of common charcoal, manifests a surprising power of whitening various substances; which, according to Mr Delaval's theory, proceeds from the power it has of dissolving the colouring matter with which they are impregnated. This solvent power, according to our author, is manifest in many other instances besides those already mentioned. Silk is whitened by the carbonated vapours of sulphur; and this operation does not appear to differ from the change effected on flowers by the same vapour. The light of the sun is found to be a necessary and essential agent in bleaching linen, wax, and various other substances; some part of the colouring matter which impairs the whiteness of these bodies not yielding to any other solvent. Red flowers are whitened by the electric spark, of whose inflammable nature we cannot entertain the least doubt; for the spark itself is a bright flame, and yields the same smell which all other carbonated matters impart. The electric spark, in like manner, changes the blue infusion of turnsole to red (n). The effects which it produces on the turnsole, and on red flowers, do not differ from each other, except in degree only. For when vegetable matter is dissolved, it is changed from blue to red; and, when farther dissolved, it is divided into particles too minute to exhibit any colour.

Solutions effected by means of phlogiston frequently are wrongly attributed to the operation of supposed acid menstrua, as several kinds of substances are capable of being dissolved indiscriminately both by acids and phlogiston. For the purpose of distinguishing, therefore, in any case between the action of the acid solvents and that of the inflammable menstrua, it is proper to examine the nature of the matter by which either of these principles are furnished. It appears from various chemical processes, that alkalis are rendered mild, and capable of crystallization, in proportion as they are united to carbuncle. The carbonated alkaline lixivium, when saturated, is perfectly mild; and by a slight evaporation is reduced to a concrete crystalline mass, which does not deliquesce or imbibe the least moisture from the air, and no longer retains any alkaline property. M. Beaume, by an elegant and ingenious experiment, has proved the presence of carbuncle in mild alkalis, and has shown that their power of crystallizing depends on their union with that principle. He heated in a silver vessel a lixivium of mild alkali, which imparted to the silver a covering or coating of inflammable matter, by which its surface was tarnished and became black. The lixivium was several times poured out of the silver vessel, and after the surface of the metal had been freed from the tarnish, the lixivium was replaced in it, and again heated, by which the tarnish was renewed; and this was repeated till the lixivium no longer communicated any stain to the silver. The causticity of the lixivium was increased in proportion as it imparted its carbuncle to the silver; and at the end of the process the alkali became perfectly caustic and incapable of crystallizing.

"From the preceding experiments (says he) it appears, that the colouring particles of flowers and leaves are soluble in acid, alkaline, and carbonated menstrua. The other parts of vegetables consist of materials similar to those which are contained in their flowers and leaves, and undergo the same changes from the same causes. Having extracted from logwood its colouring particles by repeatedly boiling it in water, the wood was thus deprived of its yellow colour, and assumed a brown hue similar to that of oak wood. Some pieces of it thus deprived of its colour were then macerated in nitric acid; and after they had undergone the action of that acid, they were washed in a sufficient quantity of water. The wood was thus reduced to whiteness."

Here our author observes, that though most authors Logwood who treat of colouring substances describe logwood as affords only of a red colour, he was never able to procure any other colour from it than yellow. It imparts yellow tincture with water and orange colours to distilled water. Other waters extract a red tinge from it by means of the alkali which they contain. These observations are also applicable to the other dyeing woods, kermes, and various other articles of the materia medica. By a similar treatment, fustic wood also lost its colouring matter, and became white.

The results of all the experiments above related are, that the colouring matter of plants does not exhibit any colour by reflection, but by transmission only; that their solid earthly substance is a white matter; and that it is the only part of vegetables which is endowed with a reflective power; that the colours of vegetables are produced by the light reflected from this white matter, and transmitted from thence through the coloured coat or covering which is formed on its surface by the colouring particles; that whenever the colouring matter is either discharged or divided by solution into particles too minute to exhibit any colour, the solid earthly substance is exposed to view, and displays that whiteness which is its distinguishing characteristic.

Mr Delaval next proceeds to examine the coloured parts of animal substances, and finds them exactly similar, with regard to the manner in which the colouring matter is... The tinctures and infusions of cochineal and of kermes yield their colours when light is transmitted through them, but show none by reflection. On diluting fresh ox gall with water, and examining it in the phials already mentioned, that part of it which was in the neck of the phial, and viewed by transmitted light, was yellow; but the anterior surface was black, and reflected no colour. Flesh derives its colour entirely from the blood, and when deprived of it, the fibres and vessels are perfectly white; as are likewise the membranes, sinews, and bones, when freed from their aqueous and volatile parts; in which case they are a mere earth, unalterable by fire, and capable of imparting an opaque whiteness to glass.

On examining blood diluted with water in one of the phials formerly described, it transmitted a red colour, and the anterior surface was almost, but not entirely, black; for it received a slight hue of brown from some coagulated particles that were suspended in the liquor. In order to procure blood sufficiently diluted, and at the same time equably and perfectly dissolved, he mixed as much cruror with spirit of sal ammoniac as imparted a bright colour to it. The liquor being then viewed in the phial, that part which was contained in the neck, and transmitted the light, appeared of a fine red; but the anterior part reflecting no light, was intensely black. Hence it appears, that the florid red colour of the flesh arises from the light which is reflected from the white fibrous substance, and transmitted back through the red transparent covering which the blood forms on every part of it.

Blood, when recently drawn, does not assume the appearance common to transparent coloured liquors; for these, when too massy to transmit light from their farther surfaces, always appear black; but blood, when recently drawn, always shows a fine red colour, in whatever way it be viewed. This is occasioned by a white matter diffused through the blood; and which is easily separated from the cruror, by dividing it after coagulation into a number of thin pieces, and washing in a sufficient quantity of pure water. Thus the water acquires a red colour, and ought to be changed daily. In a few days it will acquire no more tinge; and the remaining masses of the cruror are no longer red, but white.

In like manner, the red colour of the shells of lobsters, after boiling, is no more than a mere superficial covering spread over the white calcareous earth of which the shells are composed, and may be easily removed from the surface by scraping or filing. Before the application of heat, this superficial covering is much denser, insomuch that, in some parts of the shell, it appears quite black, being too thick to admit the passage of the light to the shell and back again; but where this transparent blue colour of the unboiled lobster is thinner, it constantly appears like a blue film. In like manner, the colours of the eggs of certain birds are entirely superficial, and may be scraped off, leaving the white calcareous earth exposed to view.

The case is the same with feathers, which owe their colours entirely to a very thin layer of some transparent matter upon a white ground. Our author ascertained this by scraping off the superficial colours from certain feathers which were strong enough to bear the operation; and thus separated the coloured layers from the white ground on which they had been naturally spread. The lateral fibres of the feathers cannot indeed have their surfaces separated in this manner; but their texture, when viewed by a microscope, seems to indicate, that the colours are produced upon them by no other means than those already related. In the examination of some animal subjects, where the colouring matter could not be separated by chemical means, our author had recourse to mechanical division; but this can only be employed when the principal part of the white substance is unmixed with the coloured coat or covering which is spread upon its surface. All of them, however, by whatever means their colours could be separated, showed that they were produced in the same manner, namely, by the transmission of light from a white ground through a transparent coloured medium.

The coloured substances of the mineral kingdom are very numerous, and belong principally to two classes, viz. earths and metals. The former, when pure, are all perfectly white, and their colours arise from carbonic or metallic mixtures. Calcareous earths, when indurated, constitute marble, and may be tinged with various colours by means of metallic solutions: all which are similar in their nature to the dyes put upon silk, cotton, or linen, and invariably proceed from the same cause, viz. the transmission of light through a very thin and transparent coloured medium. Flints are formed from siliceous earths, and owe their colour to carbonic. When sufficiently heated, they are rendered white by the loss of the inflammable matter which produced their colour. When impregnated with metals, they form agates, cornelians, jasper, and coloured crystals. The coloured gems also receive their different hues from metals: and all of them may be imitated by glasses tinged with such carbonic or metallic matters as enter into the composition of the original substances.

Thus our author concludes, that the coloured earths, gems, &c. exhibit their various tints in the same manner with other substances; viz. by the transmission of light reflected from a white ground. Our author, however, proceeds farther; and asserts, that even the colours of metals themselves are produced in the same manner.

"Gold (says he) exhibits a white light, which is tinged with yellow. I have used this expression, because it appears from experiment that gold reflects a white light, and that its yellow colour is a tinge superadded to its whiteness. The experiment is thus set forth by Sir Isaac Newton. Gold in this light (that is, a beam of white light) appears of the same yellow colour as in day light, but by intercepting at the lens a due quantity of the yellow-making rays, it will appear white like silver, as I have tried; which shows that its yellowness arises from the excess of the intercepted rays, tinging that whiteness with their colour when they are let pass.

"I have already shown, by numerous experiments, in what manner coloured tinges are produced; and it uniformly appears, from all these experiments, that colours do not arise from reflection, but from transmission only. A solution of silver is pellucid and colourless. A solution of gold transmits yellow, but reflects..." This metal also, when united with glass, yields no colour by reflection, but by transmission only. All these circumstances seem to indicate, that the yellow colour of gold arises from a yellow transparent matter, which is a constituent part of that metal; that it is equally mixed with the white particles of the gold, and transmits the light which is reflected by them, in like manner as when silver is gilt, or foils are made by covering white metals with transparent colours. But these factitious coverings are only superficial; whereas the yellow matter of gold is diffused throughout the whole substance of the metal, and appears to envelop and cover each of the white particles. In whatsoever manner the yellow matter of gold is united to its white substance, it exists in a rare state; for it bears only the same proportion to the white particles of the gold as that of the yellow-making rays which were intercepted bear to all the other rays comprised in the white light of the sun.

"Sir Isaac Newton has shown, that when spaces or interstices of bodies are replenished with media of different densities, the bodies are opaque; that those surfaces of transparent bodies reflect the greatest quantity of light which intercede media that differ most in their refractive densities; and that the reflections of very thin transparent substances are considerably stronger than those made by the same substances of a greater thickness. Hence the minute portions of air, or of the rarer medium which occupies spaces void of other matter, reflect a vivid white light whenever their surfaces are contiguous to media whose densities differ considerably from their own; so that every small mass of air, or of the rarer medium which fills the pores or interstices of dense bodies, is a minute white substance. This is manifest in the whiteness of froth, and of all pellucid colourless bodies; such as glass, crystal or salts, reduced to powder, or otherwise flaved: for in all these instances a white light is reflected from the air or rarer medium which intercede the particles of the denser substances whose interstices they occupy."

From these principles our author takes occasion to explain the reason why the particles of metals, which yield no colour by incident light when suspended in their solvents, are disposed to exhibit colours when separated from them. Hence also we see why opaque white substances are rendered pellucid by being reduced to uniform masses, whose component parts are everywhere nearly of the same density; for as all pellucid substances are rendered opaque and white by the admixture of pellucid colourless media of considerably different densities, they are again deprived of their opacity by extricating these media which kept their particles at a distance from each other; thus froth or snow, when resolved into water, lose their whiteness, and assume their former pellucid appearance. In like manner, by proper fluxes, the opaque white earths are reduced to pellucid colourless glasses; because all reflections are made at the surfaces of bodies differing in density from the ambient medium, and in the confines of equally dense media there is no reflection.

As the oxides of metals are enabled to reflect their colours by the intervention of the particles of air; so, when mixed with oil in the making of paint, they always assume a darker colour, because the excess of the density of oil over that of air forms a sensible difference when comparatively considered with respect to the specific gravity of the rarer metals. From this cause perceptibly less light is reflected from the molecule of oil than from those of air, and consequently the mass appears darker. The case, however, is different with such paints as are formed of the denser metals; as vermilion, minium, &c.; for though oil differs very considerably from air in its specific density, yet it also differs very much in this respect from the denser metallic powders: and the molecule of oil which divide their particles, act upon the light so strongly, that the reflection occasioned by them cannot be distinguished from those which are caused by rarer media. Hence though we mix vermilion or minium with oil, the colour is not less sensibly altered.

This part of our author's theory, however, seems liable to objection; for though it be true that the oxides of some metals are denser than others, yet that is not the case with all; nor is even the difference of density between oil and the oxides of the heavier metals at all comparable to that between the density of air and oil. Thus, though the oxide of iron may be 10 or 11 times more dense than oil; yet as the latter is between 500 or 600 times denser than air, the small difference between the oil and metallic oxide ought to be imperceptible. In this respect, indeed, there are considerable differences with regard to the oils employed, which cannot be supposed to arise from the mere circumstances of density. Thus the colour of vermilion, when mixed with turpentine-varnish, is much brighter than with linseed oil; and yet the difference between the densities of linseed oil and turpentine-varnish is very trifling. The mere action of heat likewise has a surprising effect in this case. Thus the red oxide of iron, called scarlet-oker, by being only heated a certain degree, appears of a very dark purple, resuming its red colour when cold; and this variation may be induced as often as we please by only heating it over the fire in a shovel. In like manner, by gradually heating red lead, it may be made to assume a most beautiful crimson colour; which growing gradually darker, becomes at last almost quite black. On cooling, if the heat has not been raised too high, it gradually returns through the same shades of colour, until at last it fixes in its original hue. These immense differences in colour cannot by any means be attributed either to the expulsion of air, or to an alteration in density. The fire indeed does certainly expand these oxides as well as other bodies; but as the medium interspersed between their particles is thus also expanded, the colour ought at least to remain the same, if not to become lighter, on account of the superior expansion of air to that of metal by the same degree of heat. It would seem, therefore, that the action of the element of fire itself has a considerable share in the production of colours; and indeed its share in the operations of nature is so great, that we might well think it strange if it should be entirely excluded from this.

With regard to semipellucid substances, which appear of one colour by incident, and another by transmitted light, our author likewise endeavours to show that no reflection is made by the coloured matter, but substances...